Intranasal Wnt3a Attenuates Neuronal Apoptosis Through Frz1/Piwil1a/FOXM1 Pathway in MCAO Rats

The Journal of Neuroscience, July 25, 2018 • 38(30):6787–6801 • 6787

Neurobiology of Disease Intranasal wnt3a Attenuates Neuronal Apoptosis through Frz1/PIWIL1a/FOXM1 Pathway in MCAO Rats

X Nathanael Matei,1 XJustin Camara,1 XDevin McBride,1,4 XRichard Camara,1 XNingbo Xu,1 X Jiping Tang,1 and X John H. Zhang1,2,3 Departments of 1Physiology and Pharmacology, 2Anesthesiology, 3Neurosurgery, Loma Linda University, Loma Linda, California 92354, and 4Vivian L. Smith Department of Neurosurgery, McGovern Medical School, University of Texas Health Science Center at Houston, Houston, Texas 77030

After ischemic stroke, apoptosis of neurons is a primary factor in determining outcome. Wnt3a is a naturally occurring protein that has been shown to have protective effects in the brain for traumatic brain injury. Although wnt3a has been investigated in the phenomena of neurogenesis, anti-apoptosis, and anti-inflammation, it has never been investigated as a therapy for stroke. We hypothesized that the potential neuroprotective agent wnt3a would reduce infarction and improve behavior following ischemic stroke by attenuating neuronal apoptosis and promoting cell survival through the Frizzled-1/PIWI1a/FOXM1 pathway in middle cerebral artery occlusion (MCAO) rats. A total of 229 Sprague Dawley rats were assigned to male, female, and 9-month-old male MCAO or sham groups followed by reperfusion 2 h after MCAO. Animals assigned to MCAO were either given wnt3a or its control. To explore the downstream signaling of wnt3a, the following interventions were given: Frizzled-1 siRNA, PIWI1a siRNA, and PIWI1a-clustered regularly interspaced short palindromic repeats, along with the appropriate controls. Post-MCAO assessments included neurobehavioral tests, infarct volume, Western blot, and immunohistochemistry. Endogenous levels of wnt3a and Frizzled-1/PIWI1a/FOXM1 were lowered after MCAO. The administration of intranasal wnt3a, 1 h after MCAO, increased PIWIL1a and FOXM1 expression through Frizzled-1, reducing brain infarction and neurological deficits at 24 and 72 h. Frizzled-1 and PIWI1a siRNAs reversed the protective effects of wnt3a after MCAO. Restoration of PIWI1a after knockdown of Frizzled-1 increased FOXM1 survival protein and reduced cleaved caspase-3 levels. In summary, wnt3a decreases neuronalapoptosisandimprovesneurologicaldeficitsthroughFrizzled-1/PIWI1a/FOXM1pathwayafterMCAOinrats.Therefore,wnt3a is a novel intranasal approach to decrease apoptosis after stroke. Key words: FOXM1; Frizzled-1; MCAO; PIWI1a; stroke; wnt3a

Significance Statement Only 5% of patients receive recombinant tissue plasminogen activator after stroke, and few qualify for mechanical thrombectomy. No neuroprotective agents have been successfully translated to promote neuronal survival in stroke. Thus, using a clinically relevant rat model of stroke, middle cerebral artery occlusion, we explored a novel intranasal administration of wnt3a. wnt3a naturally occurs in the body and crosses the blood–brain barrier, supporting the clinically translatable approach of intranasal administration.Significantneuronalapoptosisoccursduringstroke,andwnt3ashowspromiseduetoitsantiapoptoticeffects.We investigated whether wnt3a mediates its poststroke effects via Frizzled-1 and the impact on its downstream signaling molecules, PIWI1a and FOXM1, in apoptosis. Elucidating the mechanism of wnt3a will identify additional pharmacological targets and further understanding of stroke.

Introduction cause of mortality globally and the number 1 cause of disability Stroke is an acute onset disturbance of cerebral function due to (Mackay and Mensah, 2004). Two approaches have been pursued ischemia or hemorrhage. Ischemic stroke is the third leading to reduce the disease burden of stroke: neuroprotection and reperfusion (Fisher and Brott, 2003). Of these two, only reperfusion has had success in human clinical trials in the form of pharma- Received Aug. 19, 2017; revised March 27, 2018; accepted April 4, 2018. cological recombinant tissue plasminogen activator intervention Author contributions: N.M., D.M., J.T., and J.H.Z. designed research; N.M., D.M., and N.X. performed research; N.M., J.C., and R.C. analyzed data; N.M., J.C., D.M., R.C., and J.H.Z. wrote the paper. and mechanical clot disruption (Rha and Saver, 2007). Recombi- This work was supported by National Institutes of Health Grant NS081740 to J.H.Z. nant tissue plasminogen activator continues to be the only FDA- The authors declare no competing financial interests. CorrespondenceshouldbeaddressedtoDr.JohnH.Zhang,DepartmentsofAnesthesiologyandPhysiology,Loma Linda University School of Medicine, Risley Hall, Room 219, 11041 Campus Street, Loma Linda, CA 92354. E-mail: DOI:10.1523/JNEUROSCI.2352-17.2018 [email protected]. Copyright © 2018 the authors 0270-6474/18/386787-15$15.00/0 6788 • J. Neurosci., July 25, 2018 • 38(30):6787–6801 Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO approved pharmacological intervention approved for acute ischemic stroke despite multiple clinical trials exploring alternative treatments (Roth, 2011). Thus, we evaluated the potential neuroprotective agent wnt3a to protect against neuronal damage caused by ischemic stroke. wnt3a is a glycolipoprotein that controls important cellular functions, includ- ing apoptosis, neurogenesis, embryonic development, adult homeostasis, and cell polarity (MacDonald et al., 2009; Zhang et al., 2011; Wu et al., 2014). Nineteen different isoforms of “wnt” all bind to the N-terminal extracellular cysteine-rich do- main of Frizzled (Frz) family receptor, a serpentine G-protein receptor (Rao and Ku¨hl, 2010), which has its own specific isoforms. Each wnt isoform has specific innate functions dependent in part on cell type. For example, in dendritic cells, the wnt3a isoform upregulates VEGF, whereas the wnt5a isoform induces IL-10 through alternative pathways (Oderup et al., 2013). Although the wnt pathway has been studied extensively, the identifica- tion of new isoforms of wnt and their unique mechanistic endpoints have re- vived therapeutic interest in wnt (Cerpa et al., 2009; Shruster et al., 2012; Inestrosa and Varela-Nallar, 2014). Recent studies have targeted GSK-3␤ reduction to improve outcomes after ischemic stroke. In the diabetic ischemic stroke rat model, insulin, an inhibitor of the activation of GSK-3␤ (Cohen and Goedert, 2004), and TDZD-8, a selective inhibitor of GSK-3␤, both reduced the se- verity of the deleterious consequences of stroke: reducing the infarct volume and lowering the expression of GSK-3␤ (Col- lino et al., 2009). Convincingly establish- ing a role for GSK3␤ in stroke pathology, Kelly et al. (2004) showed that inhibition Figure 1. Proposed pathway for wnt3a activation of Frz1 and its downstream targets. wnt3a binds to Frizzled-1, increases of GSK3␤ with CXhir025 in both OGD in PIWIL1a, inhibits the activation of GSK3␤, and activates ␤-catenin, which translocates into the nucleus to increase the transcrip- vitro and in vivo rat middle cerebral artery tion of FOXM1, a survival protein that inhibits apoptosis. occlusion (MCAO) models increased neuronal cell survival. In the intracerebral hemorrhage mouse way in lung cancer cells, but further research is needed to estab- model, 6-bromoindirubin-3Ј-oxime, a selective GSK-3␤ inhibi- lish the relationship between PIWI1a and wnt3a in stroke. tor, acutely administered, reduced hematoma volume by attenu- ␤ ␤ Reports have linked -catenin with FOXM1 (Zhang et al., 2011), ating the expression of GSK-3 phosphorylation/activation, and both have been linked to the inhibition of apoptosis (Jiang et which increased the viability of neurons and other cell types al., 2014). ␤ (Zhao et al., 2017). These findings suggest that the wnt/GSK-3 / To date, there are 10 human G-protein-coupled Frz receptors, ␤ -catenin pathway may be further explored for treatment in the but only Frz1 has been supported in the literature to be activated pathology of stroke. by the binding of wnt3a (Chaco´n et al., 2008; Oderup et al., In this study, the novel glycolipoprotein wnt3a is used for the 2013; Wu et al., 2014; Arra´zola et al., 2015; R&D Systems, very first time as an intranasal (iN) therapy to influence the 2018). Although wnt3a is known to bind to Frz1, the intracel- pathophysiology of stroke in the MCAO model. Other studies of lular pathway triggered within neurons remains poorly inves- wnt3a have found that it prevents apoptosis (Reeves et al., 2012) tigated. In the present study, we hypothesize the effect of iN and is neuroprotective, but the mechanistic understanding is wnt3a on its downstream targets, Frz1/PIWI1a/FOXM1, will limited. Researchers showed that PIWL1a (Reeves et al., 2012) attenuate neuronal apoptosis in the rat model of transient directly affects the regulation of ␤-catenin in a cell-survival path- MCAO. (Fig. 1). Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO J. Neurosci., July 25, 2018 • 38(30):6787–6801 • 6789

Materials and Methods Intracerebroventricular injection. As described previously (Liu et al., Animals. All experiments were approved by the Institutional Animal 2007; Chen et al., 2013), rats were anesthetized with isoflurane (4% in- Care and Use Committee of Loma Linda University, complied with the duction, 2.5% maintenance) and mounted on a stereotaxic frame. The ␮ National Institutes of Health’s Guide for the care and use of laboratory needle of a 10 l Hamilton syringe was inserted through a burr hole animals, and are reported according to the ARRIVE (Animal Research: perforated through the skull into the left lateral ventricle using the fol- Reporting of In Vivo Experiments) guidelines. Animals were housed in a lowing coordinates relative to bregma: 1.5 mm posterior, 1.0 mm lateral, 12 h light/dark cycle, temperature-controlled room. Animals were di- and 3.2 mm below the horizontal plane of the bregma. siRNA was injected 24 h before MCAO. According to the manufacturer’s instructions, vided into groups in a randomized fashion, and experiments were per- ␮ formed in a blinded manner. a total volume of 2 l (500 pmol) of siRNA in sterile saline was injected in ipsilateral ventricle at a rate of 0.5 ␮l/min. A combination of two siRNAs Laboratory animals. Adult male (weighing 250–300 g), aged male (weigh- were used for Frz1 (sc-39977 and AM16708) and PIWI1a (sc-40677 and ing 400 g at 9 months of age), female (weighing 240–250 g) Sprague Dawley AM16708) providing advantages in both potency and specificity of gene rats were used in the proposed study. silencing. The same volume of scramble siRNA (sc-37007 and AM4611) Experimental design. A total of 229 rats were used. was used as a negative control. siRNA is a tool that induces short-term Experiment 1 characterized the endogenous time course of wnt3a, silencing of protein coding genes and targets a specific mRNA for degra- Frz1, PIWI1a, p-GSK3␤, FOXM1, and cleaved caspase-3 (CC-3) in the dation. To reverse the effects of siRNA, we will use an engineered form of right hemisphere at 6, 12, 24, and 72 h after MCAO. Twenty-six rats were CRISPR-associated (Cas) protein system (Jinek et al., 2012). Briefly, in used in the following groups: sham (n ϭ 4),6h(n ϭ 5), 12 h (n ϭ 6), 24 h this system, the Type II CRISPR protein Cas9 is directed to genomic (n ϭ 6), and 72 h (n ϭ 5). target sites by short RNAs, where it functions as a endonuclease, which Experiment 2 evaluated the role of wnt3a in the pathophysiology of can inactivate or activate genes (Perez-Pinera et al., 2013), used success- MCAO as follows. First, we tested the effect of iN wnt3a. A total of 203 fully in plants and animals, both vertebrae and invertebrae (Horvath and rats were used in the following groups: 24 h endpoints: sham (n ϭ 11), Barrangou, 2010; Gratz et al., 2013; Jiang et al., 2013; Mali et al., 2013; MCAO ϩ vehicle (n ϭ 14), MCAO ϩ wnt3a low dose (LD; 0.4 ␮g/kg) Bortesi and Fischer, 2015). CRISPR will be used to activate PIWI1a ex- (n ϭ 8), MCAO ϩ wnt3a high dose (HD; 1.2 ␮g/kg) (n ϭ 13), MCAO ϩ pression in the rat brain. A total of 4 ␮l of active PIWI1a CRISPR (sc- wnt3a ϩ scrambled siRNA (n ϭ 7), MCAO ϩ wnta3a ϩ Frz1-siRNA 418611) was injected 24 h before MCAO; 20 ␮g of CRISPR was (n ϭ 8), MCAO ϩ wnt3a ϩ PIWI-siRNA (n ϭ 8), MCAO ϩ wnt3a ϩ suspended in 20 ␮l of plasmid transfection medium (sc-108062) and scrambled clustered regularly interspaced short palindromic repeats ␮ ϭ ϩ ϩ ϩ ϭ activated with another 20 l of transfection reagent (sc-395739), totaling 2 (CRISPR; n 7), MCAO wnt3a PIWI CRISPR Frz1-siRNA (n ␮g per animal of active CRISPR. The control scrambled CRISPR (sc- 7); female: sham (n ϭ 6), MCAO ϩ vehicle (n ϭ 8), MCAO ϩ wnt3a HD ␮ ␮ ϭ ϭ ϩ ϭ 437275) followed the same steps, and a total of 2 g per animal was given (1.2 g/kg) (n 8); aged rats: sham (n 6), MCAO vehicle (n 7), intracerebroventricularly. To prevent possible leakage, the needle was ϩ ␮ ϭ MCAO wnt3a HD (1.2 g/kg) (n 6); permanent MCAO (pMCAO): kept in situ for an additional 10 min after completing the injection and ϩ ϭ ϩ ␮ ϭ MCAO vehicle (n 10), MCAO wnt3a HD (1.2 g/kg) (n 9); then withdrawn slowly over 5 min. After removal of the needle, the burr ϭ ϩ ϭ ϩ Naive groups: Naive (n 6), Naive scrambled siRNA (n 12), Naive hole was sealed with bone wax, the incision was sutured, and the rats were ϭ ϩ ϭ ϩ Frz1-siRNA (n 6), Naive PIWI-siRNA (n 6), Naive PIWI1a allowed to recover. ϩ ϭ ϭ ϩ CRISPR Frz-1siRNA (n 6); 72 h endpoints: Sham (n 6), MCAO 2,3,5-Triphenyltetraolium chloride staining. As described previously ϭ ϩ ϭ vehicle (n 8), MCAO wnt3a HD (n 8). Recombinant mouse (Hu et al., 2009), 2,3,5-triphenyltetrazolium chloride monohydrate wnt3a (Abcam; ab81484), used for all molecular and behavioral studies, staining was performed to determine the infarct volume at 24 and 72 h and human GST tagged wnt3a (Abcam; 153563), only used in Figure 13 after MCAO. The possible interference of brain edema with infarct volume to confirm the presence and delivery, were reconstituted in ddH2Oto was corrected (whole contralateral hemisphere volume minus nonischemic ␮ ␮ yield 0.1 mg/ml and given at 0.4 g/kg for the LD or 1.2 g/kg for the HD. ipsilateral hemisphere volume) and the infarcted volume was expressed as a A total of 500 pmol of Frz1 (sc-39977 and AM16708) and PIWI1a (sc- percentage of the whole contralateral hemisphere (McBride et al., 2016). ␮ 40677 and AM16708) was dissolved in 2 l of sterilized water and in- Immunofluorescence staining. Immunohistochemistry was performed jected intracerebroventricularly 24 h before MCAO. The same volume of as described previously (Chen et al., 2013). Briefly, rats were perfused scramble siRNA (sc-37007 and AM4611) was administered as control. with cold PBS under deep anesthesia, followed by infusion of 4% forma- ␮ ␮ For PIWI1a-CRISPR activation, 2 gofin2 l (transfection medium lin 24 h after MCAO. The brains were harvested and immersed in 4% and transfection reagent, described in Intracerebroventricular injection) formalin at 4°C, then dehydrated with 30% sucrose for 7 d, then mounted per animal was injected 24 h before MCAO (detailed in Intracerebroven- in OCT and frozen. After cryosectioning into 10-␮m-thick sections, slices tricular injection). The control of a scrambled CRISPR was used with the were incubated with the primary antibodies goat anti-Frz1 (1:100) (PA5– same volume. 47072, RRID:AB_2610121), rabbit anti-NeuN (1:200) (ab177487, RRID: MCAO model. Rats were anesthetized intraperitoneally with a mixture AB_2532109), rabbit anti-GFAP (1:100) (ab16997, RRID:AB_443592), of ketamine (80 mg/kg) and xylazine (20 mg/kg), until no pinch-paw goat anti-ionized calcium binding adaptor molecule 1 (IBA1, 1:200; reflex was observed, and maintained throughout the experiments. Body ab107159, RRID:AB_10972670), followed by incubation with appropri- temperature and respiration rate were monitored perioperatively. A ver- ate secondary antibodies conjugated with either FITC (Neun, GFAP, tical midline cervical incision was made, and the right common carotid Iba1) or Rodamine Red (Frz1) and DAPI (nuclear marker, color blue) artery was exposed and dissected. All branches of the external carotid (Jackson ImmunoResearch Laboratories). Negative control staining was artery were isolated, coagulated, and transected, and then the external performed by omitting the primary antibody. The sections were visual- carotid artery was divided (leaving 3–4 mm). The internal carotid artery ized with a fluorescence microscope (Leica Microsystems). was isolated, and the pterygopalatine artery was ligated close to its origin. Western blots. Western blotting was performed as described previously A 5 mm aneurysm clip was used to clamp the common carotid artery. (Ayer et al., 2012; Chen et al., 2013). At each time point, rats were per- The external carotid artery stump was reopened, and a 4.0 monofilament fused with cold PBS, pH 7.4, solution delivered via intracardiac injection, nylon suture with an enlarged, round tip was inserted through the inter- followed by dissection of the brain into right frontal, left frontal, right nal carotid artery; insertion was stopped when resistance was felt, occlud- parietal, left parietal, cerebellum, and brainstem. The brain parts were ing the origin of the right MCA. After 2 h of occlusion, the suture was snap frozen in liquid nitrogen and stored at Ϫ80°C for subsequent anal- withdrawn to allow for reperfusion. The external carotid artery was li- ysis. Right hemisphere protein extraction from whole-cell lysates were gated, and the aneurysm clip was removed. The skin was sutured (1% obtained by gently homogenizing the hemisphere in RIPA lysis buffer lidocaine was applied), and the animal was allowed to recover. Sham (Santa Cruz Biotechnology) with further centrifugation at 14,000 ϫ g at surgery included the exposure of the common, external, and internal 4°C for 30 min. The supernatant was used as whole-cell protein extract, carotid arteries with all ligations and transections; no occlusion occurred and the protein concentration was determined using a detergent- in the sham group (Kusaka et al., 2004; Chen et al., 2011). compatible assay (Bio-Rad). Equal amounts of protein were loaded on an 6790 • J. Neurosci., July 25, 2018 • 38(30):6787–6801 Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO

Figure2. Timecourseofwnt3a,Frz1,PIWI1a,pGsk-3␤,FOXM1,andCC-3intherighthemisphereoftheratbrainafterMCAO.A,RepresentativeWesternblots.QuantitativeanalysiswithWestern blot showed that the expression of (B) wnt3a and (C) Frz1 significantly decreased by 6 h. D, PIWI1a was decreased by 12 h and returned to sham levels by 72 h. E, However, pGsk-3␤ was increased after MCAO. F, FOXM1 was decreased after MCAO, but (G) CC-3 was increased. Each column represents the mean Ϯ SD (n ϭ 4/group). *p Ͻ 0.05 versus Sham. #p Ͻ 0.05 versus 6 h. Figure 2-1 (available at https://doi.org/10.1523/JNEUROSCI.2352-17.2018.f2-1) reports the specific statistics for each group comparison. Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO J. Neurosci., July 25, 2018 • 38(30):6787–6801 • 6791

The experimenter holds the animal so that all four limbs hang freely, and the vibrissae are stimulated by sweeping each side against the edge of a table. This elicits an ipsilateral forelimb response to place the paw on the table top. The rat is stimulated 10 times on each side, and the total number of paw placements is recorded. Fluro-Jade-C (FJC) staining. FJC staining was performed to detect degenerating neurons with a modified FJC ready-to-dilute staining kit (Biosensis) according to the manufacturer’s instructions (Schmued et al., 2005). FJC-positive neurons in a 0.56 mm 2 region of the ischemic boundary zone next to the ischemic core were counted in four randomly selected mi- croscopic fields by an independent observer. Quantified analysis was performed by blinded researchers using ImageJ software. The data were presented as the number of FJC-positive neurons in the FOV. Statistical analysis. Quantitative data are presented as mean Ϯ SD. One-way ANOVA with post hoc Tukey test was used to determine significant group differences among groups at each time point for infarction volume and West- ern blot data. For nonparametric data, one-way ANOVA-Wallis with Dunn’s post hoc was used to analyze neurobehavioral data after deviation from normality was confirmed. p Ͻ 0.05 was considered statistically significant. Prism 6 was used for graphing and analyzing all data Figure 3. wnt3a attenuated infarct volume and improved neurological function 24 h after MCAO. A, Representative triphenyl- (GraphPad). tetrazolium chloride (TTC) staining images of coronal sections. B, wnt3a HD (1.2 ␮g/kg) effectively reduced the infarct volume. C, ModifiedGarciascoresshowedthatLDandHDofwnt3adecreasedneurologicaldeficits.D,Leftforelimbplacementwasimproved Results Ϯ ϭ Ͻ # Ͻ in both the LD and HD of wnt3a. Each column represents the mean SD (n 6–9/group). *p 0.05 versus Sham. p 0.05 All sham-operated rats survived, but the versus vehicle. overall mortality of MCAO was 21.94%. The mortality was not significantly differ- SDS-PAGE gel. After being electrophoresed and transferred to a nitro- ent among the experimental groups (data not shown). cellulose membrane, the membrane was blocked and incubated with the primary antibody overnight at 4°C. The primary antibodies were rabbit Endogenous expressions of wnt3a and FRZ1/PIWI1/FOXM1 polyclonal anti-wnt3a (1:1000; ab28472, RRID:AB_2215308), rabbit polyclonal were decreased in neurons 24 h after MCAO, while the anti-Frz1 (1:1000; ab126262, RRID:AB_11128657), rabbit polyclonal anti- ␤ PIWI1a (1:1000; ab12337, RRID:AB_470241), mouse monoclonal proteins Caspase 3 and GSK3- were increased anti-␤-catenin (1:1000; ab32572, RRID:AB_725966), rabbit poly- Western blot was used to detect time course expression of wnt3a, clonal anti-pGsk-3␤ (1:1000; ab75745, RRID:AB_1310290), rabbit poly- Frzd1, PIWI1a, FOXM1, pGSK-3␤, and CC-3 in brain tissue at clonal anti-FOXM1 (1:1000; sc-376471, RRID:AB_11150135), rabbit baseline, 6, 12, 24, and 72 h after MCAO (Fig. 2). Representatives polyclonal anti-CC-3 (1:1000; ab13847, RRID:AB_443014), rabbit polyclonal immunoblots are shown in Figure 2A. The expression pattern of anti-GST (ab19256, RRID:AB_444809), rabbit monoclonal anti-lipoprotein wnt3a (Fig. 2B) and Frz1 (Fig. 2C) showed a significant decrease receptor-related protein-6 (LRP6) (1:1000; ab134146), and rabbit polyclonal from 6 to 72 h after MCAO (Fig. 2B). PIWI1a was significantly anti-phospho-LRP6 (1:500; 2568s, RRID:AB_2139327). For loading control, decreased from 12 to 24 h after MCAO, but indistinguishable at ␤ the same membranes were blotted with primary antibody of goat anti- - 72 h compared with sham (Fig. 2D). After MCAO, p-GSK3␤ actin (1:1000; sc-1616, RRID:AB_630836). Nitrocellulose membranes increased significantly from 12 to 72 h compared with sham (Fig. were incubated with appropriate secondary antibodies (1:2000; Santa Cruz Biotechnology) for 30 min at room temperature. Immunoblot 2E). Survival protein FOXM1 was significantly decreased from 6 to bands were further probed with a chemiluminescence reagent kit (ECL 72 h after MCAO (Fig. 2F). In contrast, apoptotic protein CC-3 was Plus; GE Healthcare). Data were analyzed by blinded researchers using significantly increased from 6 to 72 h after MCAO (Fig. 2G). Figure ImageJ software, and each protein was normalized to their respective 2-1 (available at https://doi.org/10.1523/JNEUROSCI.2352-17.2018. ␤-actin bands. f2-1) reports the specific statistics for each group comparison. Neurobehavioral testing. Neurobehavioral outcomes were assessed by a blinded observer at 24 and 72 h after MCAO using the modified Garcia Effects of wnt3a on infarction size and neurobehavioral Score (Kusaka et al., 2004; Chen et al., 2010). Animal scores for sensori- function 24 h after MCAO motor functions evaluated six parameters: spontaneous activity, symme- Qualitative images of infarction volume in various groups are try in the movement of all four limbs, forepaw outstretching, climbing, shown in Figure 3A; and quantitively, there was a significant body proprioception, and response to vibrissae touch. A maximum score ␮ of 21 was given with higher scores indicating better performance. reduction in infarction volume with high dosage (1.3 g/kg) Left forelimb testing. Vibrissae-evoked forelimb placing test was as- compared with vehicle (one-way ANOVA; Tukey’s test; n ϭ 6–9; ϭ ϭ sessed by a blinded observer at 24 and 72 h after MCAO. Left forelimb F(3,20) 23.44; p 0.0153) (Fig. 3B). wnt3a restored modified testing assesses for asymmetry in the sensorimotor cortex and striatum. Garcia (Fig. 3C) scores to sham levels and caused significant im- 6792 • J. Neurosci., July 25, 2018 • 38(30):6787–6801 Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO provement in left forelimb placement (Fig. 3D) compared with vehicle (one-way ANOVA; Kruskal–Wallis with Dunn’s post hoc; n ϭ 6–9; p ϭ 0.0031) at 24 h. With the loss of oxygen and glucose, a hypoxic state is created in the acute stage of cerebral infarction, and recanalization triggers an apoptotic neuronal stress response (Candelario-Jalil, 2009). Survival of neurons is a key contributor to outcomes after stroke (Lai et al., 2014). Thus, neuroprotection is an important therapeutic target for the treatment of stroke (Cerpa et al., 2009; Lai et al., 2014). To evaluate the effects of iN wnt3a on neurological outcomes in transient focal cerebral ischemia, infarction volume and neurobehavior were assessed. wnt3a reduced apoptotic cells after MCAO Brain coronal sections obtained 24 h after MCAO were stained with FJC to visualize and count degenerating cells in the penumbra of the cortex after ischemia/reperfusion (Fig. 4A). Higher levels of FJC staining were observed in the MCAO group compared with the Sham group. wnt3a significantly decreased the FJC- positive cells in the penumbra compared with vehicle (67.00 Ϯ 14.24/field vs 193 Ϯ 30.84/field, one-way ANOVA; Tukey’s test; ϭ ϭ Ͻ F(2,6) 72.44; n 3; p 0.05, for each) (Fig. 4B). Figure 4. The effects of wnt3a on neuronal damage at 24 h after MCAO. A, Representative microphotographs of FJC-positive Immunofluorescence of the receptor neurons. B, Quantitative analysis of FJC-positive cells was performed in the penumbra of the cortex. C, FJC-positive neurons Frz1 in the penumbra at 24 h significantly increased after MCAO induction. wnt3a significantly reduced the number of FJC-positive cells compared with vehicle Staining at the edge of the infarction (Fig. (p Ͻ 0.05). Data are mean Ϯ SD. Scale bar, 50 ␮m. n ϭ 3 per group. *p Ͻ 0.05 versus Sham. #p Ͻ 0.05 versus vehicle. 5A) at 24 h probed neurons (NeuN), astrocytes (GFAP), and microglia (Iba1). In reperfusion, Western blot was performed on the right hemi- the sham group, immunoreactivity of Frz1 (red) was present in sphere of the brain, representative images in Figure 6A.Inthe NeuN and GFAP, but not Iba-1 (Fig. 5B). Vehicle expression of 24 h MCAO ϩ vehicle group, proteins wnt3a, Frz1, PIWI1a, and Frz1 in Neun, GFAP, or Iba-1 could not be detected after MCAO. ␮ FOXM1 were significantly decreased compared with the sham HD wnt3a treatment (1.2 g/kg) demonstrated strong immuno- group (p Ͻ 0.05; Fig. 6B–D,G), whereas p-GSK3␤ and CC-3 were reactivity of Frz1 colocalized with NeuN (Fig. 5B). Although the significantly increased compared with the sham group (p Ͻ 0.05; literature agrees with Iba-1 expression on microglial cells, our obser- Fig. 6E,H). Additionally, at 24 h, HD wnt3a (1.2 ␮g/kg) iN treat- vations suggest that the expression of Frz1 on Iba-1 may need to be ment after MCAO significantly upregulated the expression of evaluated at later time points, peaking from 72 h to 7 d after Frz1, ␤-catenin, PIWI1a, and FOXM1 compared with vehicle MCAO (Michalski et al., 2012; Taylor and Sansing, 2013; Kawa- Ͻ ␤ bori et al., 2015). groups (p 0.05; Fig. 6B–D,F,G), whereas p-GSK3 and CC-3 were significantly attenuated compared with the vehicle group Ͻ ϩ Effects of wnt3a on Frz1, p-GSK3␤, ␤-catenin, PIWI1a, (p 0.05; Fig. 6E,H). The 72 h MCAO vehicle group wnt3a, FOXM1, CC-3 pathway at 24 and 72 h after MCAO Frz1, and FOXM1 were significantly decreased compared with Ͻ ␤ ␤ Of particular interest, recent studies have shown activation of sham (p 0.05; Fig. 6B,C,G), whereas p-GSK3 , -catenin, and extrinsic and intrinsic pathways of caspase-mediated cell death in CC-3 were significantly increased compared with the sham group Ͻ ϩ several forms of transient MCAO in adult rats (Ferrer and Planas, (p 0.05; Fig. 6E,F,H). In the 72 h MCAO iN wnt3a (1.2 2003). CC-3 is upregulated following MCAO, and subsequently, ␮g/kg) group, wnt3a, Frz1, ␤-catenin, and FOXM1 levels were sig- the inhibition of caspase-3 reduces infarct size following transient nificantly increased compared with vehicle groups (p Ͻ 0.05; Fig. MCAO (Ferrer and Planas, 2003). Thus, our hypothesis is that 6B,C,F,G); there was no change in PIWI1a levels at 72 h. wnt3a wnt3a will provide therapeutic benefits following ischemic stroke expression was higher in vehicle groups at 24 and 72 h compared in rats via attenuation of CC-3 and prevention of neuronal apo- with endogenous expression levels after MCAO in Figure 2B, ptosis. To investigate the effects of Frz1 activation with the spe- which could be explained by two key observations. First, with a cific HD wnt3a (1.2 ␮g/kg) agonist administered iN at 1 h after sample size of 4 in Figure 2B, the protein expression levels pro- Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO J. Neurosci., July 25, 2018 • 38(30):6787–6801 • 6793

7C), and corner-turn (Fig. 7D) scores to sham levels. In the 72 h MCAO ϩ vehicle group, infarction volume was significantly increased and neurobehavior scores for modified Garcia (one-way ANOVA; Kruskal– Wallis with Dunn’s post hoc; p ϭ 0.023; Fig. 7B), left forelimb placement (one- ϭ way ANOVA; Tukey’s test; F(2,15) 8.59; n ϭ 6; p ϭ 0.003; Fig. 7C), and corner- turn (one-way ANOVA; Tukey’s test; ϭ ϭ ϭ F(2,15) 5.86; n 6; p 0.0102; Fig. 7D) were significantly decreased compared with sham.

Impact of Frz1-siRNA, PIWI1a-siRNA, and PIWI CRISPR on the function of the wnt3a/Frz1/PIWI/FOXM1 pathway at 24 h after MCAO To elucidate the molecular mechanisms modulated by wnt3a downstream of its receptor Frz1, two inhibitor groups were used: Frz1-siRNA and PIWI siRNA. At 24 h, MCAO ϩ wnt3a ϩ Frz1-siRNA significantly decreased the protein expression of wnt3a, Frz1, PIWI1a, and FOXM1 compared with the wnt3a-treated group (p Ͻ 0.05; Fig. 8B–D,F); pGSK3␤ and CC-3 were significantly increased compared with the wnt3a-treated and scrambled control groups (p Ͻ 0.05; Fig. 8E,G). In the MCAO ϩ wnt3a ϩ PIWI siRNA group, Frz1, PIWI1a, and FOXM1 were significantly decreased compared with the treated and scramble control groups (p Ͻ Figure 5. Expression of Frz1 (red) in neurons (NeuN, green), astrocytes (GFAP, green), and microglia (Iba1, green) in sham,ve- ␤ hicle, and treatment. A, Representative triphenyl tetrazolium chloride stained coronal section showing the location of all immu- 0.05; Fig. 8C,D,F); wnt3a, pGSK3 , and nohistochemistry images indicated by the boxed region. B, Expression of Frz1 decreased in vehicle at 24 HR after MCAO in neurons CC-3 were significantly increased com- butwasincreasedwithtreatment.ShamshowedminimalexpressionofFrz1inastrocytesandnoexpressioninmicroglia,whereas pared with the treated and scramble con- vehicle and HD treatment (1.2 ug/kg) had no Frz1 expression in astrocytes and microglia (n ϭ 1 per each group). Blue represents trol groups (p Ͻ 0.05; Fig. 8B,E,G). In the DAPI. Scale bar, 50 ␮m. MCAO ϩ wnt3a ϩ Frz1-siRNA ϩ PIWI CRISPR group, protein expression of PIWI1a and FOXM1 was significantly in- vided statistical significance to merit further mechanistic investi- creased compared with the vehicle group (p Ͻ 0.05; Fig. 8D,F); gation into the significance of the pathway. Nonetheless, there wnt3a, Frz1, pGSK3␤, and CC-3 were significantly decreased com- still remains an opportunity for rather marked variation among pared with wnt3a-treated group (p Ͻ 0.05; Fig. 8B,C,E,G). Figure 8-1 small sample sizes despite statistically significant comparisons. (available at https://doi.org/10.1523/JNEUROSCI.2352-17.2018. With an increased sample size of 6, a more quantitative evaluation of f8-1) reports the specific statistics for each group comparison. the model and wnt3a expression could be observed. Second, the vehicle solution may have slightly increased the expression of endog- Sexually dimorphic comparisons of infarct size and enous wnt3a, but this expression was still significantly lower com- neurobehavioral function 24 h after MCAO in female rats pared with sham. Additionally, iN wnt3a (1.2 ␮g/kg) significantly To evaluate for sexually dimorphic differences after treatment, decreased levels of p-GSK3␤ and CC-3 compared with the vehicle infarct and neurobehavior were assessed at 24 h in female rats. group (p Ͻ 0.05; Fig. 6E,H). Figure 6-1 (available at https://doi.org/ With administration of wnt3a, infarction volume was signifi- 10.1523/JNEUROSCI.2352-17.2018.f6-1) reports the specific sta- cantly reduced compared with vehicle (one-way ANOVA; Tukey’s ϭ ϭ ϭ tistics for each group comparison. test; F(2,15) 53.11; n 6, p 0.0079) (Fig. 9B). wnt3a restored modified Garcia (Fig. 9C) and left forelimb placement (Fig. 9D) Effects of wnt3a on infarct size and neurobehavioral function scores to sham levels. HD wnt3a (1.2 ␮g/kg) treatment after 72 h after MCAO MCAO had significantly therapeutic effects in both sexes, with no To evaluate the effects of iN wnt3a HD (1.2 ␮g/kg) on long-term major intersex differences in infarct and behavioral evaluations. neurological outcome, infarct and neurobehavior were assessed at 72 h. With administration of wnt3a, infarction volume was Age-related effects on infarct size and neurobehavioral significantly reduced compared with vehicle (one-way ANOVA; function 24 h after MCAO ϭ ϭ ϭ Tukey’s test; F(2,15) 25.20; n 6; p 0.0084) (Fig. 7A). wnt3a To assess the age-related effects after MCAO, neurological outcome, restored modified Garcia (Fig. 7B), left forelimb placement (Fig. infarct, and neurobehavior were assessed at 24 h in 9-month-old 6794 • J. Neurosci., July 25, 2018 • 38(30):6787–6801 Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO

male rats. With administration of wnt3a, infarction volume was significantly reduced compared with vehicle (one-way ϭ ϭ ANOVA; Tukey’s test; F(2,15) 115.8; n 6; p ϭ 0.0124) (Fig. 10B). wnt3a had no effect on modified Garcia (Fig. 10C), and both vehicle (one-way ANOVA; Kruskal– Wallis with Dunn’s post hoc; n ϭ 6; p ϭ 0.0053) and treatment (one-way ANOVA; Kruskal–Wallis with Dunn’s post hoc; n ϭ 6; p ϭ 0.0181) modified Garcia scores were decreased compared with sham. wnt3a restored left forelimb placement (Fig. 10D) scores to sham levels. Overall, HD wnt3a (1.2 ␮g/kg) treatment in aged rats was not as advantageous compared with younger-treated MCAO cohorts; however, significant reduction of infarction and improved left forelimb placement was observed.

wnt3a effects on infarct size and neurobehavioral function at 24 h in a pMCAO model To evaluate the model variation effects on neurological outcome, infarct and neurobehavior were assessed at 24 h in a pMCAO model. With administration of wnt3a, no significant improvement was seen in infarction compared with vehicle (one-way ANOVA; Tukey’s test; ϭ ϭ ϭ F(2,15) 59.35; n 6, p 0.8816; Fig. 11B). Both modified Garcia (Fig. 11C) and left forelimb (Fig. 11D) placement showed no significant difference in treatment compared with vehicle (one-way ANOVA; Kruskal–Wallis with Dunn’s post hoc; n ϭ 6).

Impact of Frz1-siRNA, PIWI1a-siRNA, and PIWI CRISPR in naive animals Naive animals were used to evaluate the effectiveness of siRNAs and CRISPR. Na- ive ϩ Frz siRNA significantly reduced Frz-1 protein levels compared with both Naive and Naive ϩ Scrambled siRNA ϭ (one-way ANOVA; Tukey’s test; F(2,15) 34.15; n ϭ 6, p Ͻ 0.05; Fig. 12A). In the Naive ϩ PIWI1a siRNA group, PIWI1a levels were significantly attenuated compared with Naive and Naive ϩ Scrambled siRNA (one-way ANOVA; Tukey’s test;

vehicle groups but reduced by the wnt3a treatment. F, ␤-Catenin was increased by wnt3a at both time points. G,FOXM1wasbroughtbacktoshamlevelsbyHDwnt3atreat- ment, and (H) CC-3 was therefore reversed in the treatment groups. Each column represents the mean Ϯ SD (n ϭ 6/group). *p Ͻ 0.05 versus Sham. #p Ͻ 0.05 versus vehicle. Figure 6. wnt3a prevented apoptosis through Frz1/PIWI1a/␤-catenin/FOXM1 pathway after MCAO in rats. A, Right hemi- Figure 6-1 (available at https://doi.org/10.1523/JNEUROSCI. sphererepresentativeWesternblotsoftheratbrain.wnt3a(B)andFrz1(C)wereincreasedafterHDtreatment(1.2␮g/kg)inboth 2352-17.2018.f6-1) reports the specific statistics for each group 24and72hendpoints.D,PIWI1awasincreasedbywnt3aat24HRbutnotlostitssignificanceat72h.E,pGsk-3␤wasincreasedin comparison. Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO J. Neurosci., July 25, 2018 • 38(30):6787–6801 • 6795

Figure 7. Wnt3a decreased infarct volume and improved neurological function 72 hours after MCAO. A, Cerebral infarct quantification was carried out on the TTC-stained images of coronal sections,(B)modifiedGarcianeuroscore,(C)leftforelimbplacement,(D)cornerturntestshowedthathighdosageofwnt3adecreasesinfarctionandneurologicaldeficits72hoursafterMCAO.Each column represents the mean Ϯ SD (n ϭ 6/group). *p Ͻ 0.05 vs. Sham, #p Ͻ 0.05 vs vehicle.

ϭ ϭ Ͻ F(3,20) 24.51; n 6, p 0.05; Fig. 12B). This effect was reversed is needed to thoroughly evaluate the pharmacokinetics of exoge- and restored to sham levels in the Naive ϩ PIWI1a CRISPR ϩ nous wnt3a administered after MCAO. Better pharmacokinetic ϭ Frz1siRNA group (one-way ANOVA; Tukey’s test; F(3,20) 34.15; understanding will be essential in the development of clinical n ϭ 6, p Ͻ 0.05; Fig. 12B). applications and strengthening of wnt3a’s translational impact.

Evaluation of exogenous recombinant wnt3a in wnt3a and alternative coreceptor LRP6 binding treated animals When wnt proteins bind to Frz cell surface receptors, the trans- Because the recombinant mouse wnt3a (ab81484) protein was membrane LRP5/LRP6 are recruited; and in the canonical wnt used for all molecular and behavioral studies, due to its homology pathway, cytoplasmic protein Dishevelled (Dvl) is activated, to rat wnt3a, and could not be distinguished from the endogenously which leads to the inhibition of GSK3 and prevention of the expressed wnt3a, a human recombinant wnt3a (ab 153563) protein phosphorylation, degradation of ␤-catenin (Bhanot et al., 1996; with a specific tag was iN administered, 1.2 ␮g/kg, 1 h after re- Abe et al., 2013). Therefore, we further investigated the regula- canalization to check delivery and presence of our drug. The tion of LRP6 after wnt3a treatment in MCAO. iN wnt3a (1.2 human recombinant wnt3a protein was made by the parasite ␮g/kg) was administered at 1 h after reperfusion in treatment Schistosoma japonicum and had GST3 bound to the N-terminus groups, and Western blot was performed on the right hemisphere to distinguish it from endogenous proteins expressed by the rat. A of the brain, representative images in Figure 14. LRP6 levels were specific GST antibody without binding affinity for the rat gluta- not significantly different between: sham, vehicle, MCAO ϩ wnt3a, thione S-transferase P protein (GSTP1 also known as GST3) was MCAO ϩ wnt3a ϩ scrambled-siRNA, and MCAO ϩ wnt3a ϩ used to assess the amount of recombinant protein (ab153563) in Frz1-siRNA groups. However, the phosphorylated LRP6 levels our treated animals. Representative Western blots and quantita- were significantly higher in the HD (1.2 ␮g/kg) wnt3a groups ϭ tive analysis showed that the amount of recombinant wnt3a compared with vehicle (one-way ANOVA; Tukey’s test; F(4,25) GST-specific protein was significantly higher compared with 17.76; n ϭ 6, p Ͻ 0.05; Fig. 14). This effect was reversed, signifi- nontreated groups, both sham and vehicle (one-way ANOVA; cantly attenuating the pLRP6 expression in the MCAO ϩ wnt3a ϩ ϭ ϭ Ͻ ϩ Tukey’s test; F(2,15) 285.4; n 6, p 0.05; Fig. 13). Although Frz1siRNA group compared with MCAO wnt3a (one-way ϭ ϭ Ͻ exogenous delivery was confirmed in our study, further research ANOVA; Tukey’s test; F(4,25) 17.76; n 6, p 0.05; Fig. 14). 6796 • J. Neurosci., July 25, 2018 • 38(30):6787–6801 Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO

Figure8. Effectofwnt3apreventingapoptosisthroughtheFrz1/PIWI1a/FOXM1pathwayat24hafterMCAO.A,RighthemisphererepresentativeWesternblotsoftheratbrainforthepathway.B,AftersiRNAtreatment ofFrz1,wnt3areturnedtovehiclelevels,and(C)Frz1expressionwasdependentonthepresenceofwnt3a,whichwassignificantlyreducedwithFrz1-siRNAandPIWI-siRNA.WiththeincreasedexpressionofPIWI1a(D)through treatmentorCRISPR,pGsk-3␤(E)wasdecreased.F,FOXM1wasincreasedinexpressionthroughtreatmentanddecreasedthroughFrz1-siRNAandPIWI-siRNA.G,CC-3wasdecreasedaftertreatmentbutrestoredtovehicle levelsbytheinhibitionoftheFrz1/PIWI1a/FOXM1pathwaywithbothFrz1-siRNAandPIWI-siRNA.EachcolumnrepresentsthemeanϮSD(nϭ6/group).*pϽ0.05versusSham. #pϽ0.05versusvehicle. &pϽ0.05versus wnt3a.Figure8-1(availableathttps://doi.org/10.1523/JNEUROSCI.2352-17.2018.f8-1)reportsthespecificstatisticsforeachgroupcomparison.

This suggests that the phosphorylation of LRP6 is dependent on wnt3a-Frz1 binding, which is supported in the literature: wnt3a has a 2–3ϫ stronger affinity toward Frz1 compared with LRP6 (Bourhis et al., 2010). Phosphorylation of LRP6 may be responsible for additional regulation of GSK3 in our proposed neuronal pathway. Discussion In the present study, we made the following observations: (1) endogenous wnt3a and its downstream targets, Frz1/PIWI1a/FOXM1, were significantly decreased, whereas pGSK3␤ and CC-3 were significantly increased in the brain at 24 h after MCAO; (2) iN wnt3a decreased infarct volume and improved neurobehavioral function at 24 h after MCAO; (3) immunofluorescence confirmed that neurons express the receptor of wnt3a, Frz1; (4) exogenous iN wnt3a administered 1 h after MCAO significantly upregulated the expression of Frz1, ␤-catenin, PIWI1a, and FOXM1 compared with vehicle groups and decreased CC-3 levels; (5) 72 h after MCAO, infarct size and neurobehavioral function Figure 9. wnt3a attenuated infarct volume and improved neurological function 24 h after MCAO in female rats. A, Represen- improved after a single dose of wnt3a; (6) tativeTTCstainingimagesofcoronalsectionsinfemalerats.B,wnt3aHDeffectivelyreducedtheinfarctvolume.C,ModifiedGarcia specific siRNAs showed links between showed that wnt3a improved neurological function. D, Left forelimb placement was improved with wnt3a. Each column repre- Frz1, PIWI1a, and FOXM1 at 24 h after sents the mean Ϯ SD (n ϭ 6/group). *p Ͻ 0.05 versus Sham. #p Ͻ 0.05 versus vehicle. Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO J. Neurosci., July 25, 2018 • 38(30):6787–6801 • 6797

MCAO; (7) the same efficacy of wnt3a after stroke was seen in female rats, but the effect was diminished in aged rats; and (8) in the pMCAO model, no significant difference was observed between vehicle and wnt3a groups. wnt3a caused significant increase in PIWI1a and FOXM1 and a decrease in CC-3 at 24 h after wnt3a administration. At 72 h, in the vehicle group, PIWI1a levels were restored to prestroke levels. This result suggests that either the modest recovery of endogenous wnt3a levels at 72 h or an alternate pathway is responsible for the upregulation of ␤-catenin, PIWI1a, and FOXM1 levels 72 h after MCAO. In the literature, although ␤-catenin expression was endogenously elevated at 72 h after subarachnoid hemorrhage (SAH), neuronal survival was not significantly improved (Chen et al., 2015). In this study, the higher ␤-catenin expression after wnt3a treatment significantly reduced apoptosis at 24 and 72 h endpoints. To confirm that PIWI1a regulates FOXM1, a specific inhibitor of PIWI1a, Figure 10. Nine-month-old aged rats were evaluated at 24 h after MCAO and showed a reduction in infarct volume with an PIWI1a-siRNA, was administered with improved left forelimb placement after wnt3a administration. A, Representative TTC staining images of coronal sections in aged wnt3a. This intervention caused a signifi- rats. B, wnt3a effectively reduced infarct volume compared with vehicle. C, Modified Garcia was not able to detect an effect after cant decrease in FOXM1 expression at 24 h. wnt3a administration compared with vehicle. D, Left forelimb placement was improved with wnt3a. Each column represents the To further investigate the possibility of an Ϯ ϭ Ͻ # Ͻ mean SD (n 6/group). *p 0.05 versus Sham. p 0.05 versus vehicle. alternative pathway, two groups were added: wnt3a ϩFrz1-siRNA ϩMCAO and wnt3a ϩ Frz1-siRNA ϩPIWI1a-CRISPR-activation ϩ MCAO. In the cohort of wnt3a ϩ Frz1- siRNA ϩ MCAO, we saw a significant reduction of PIWI1a and FOXM1; however, in wnt3a ϩ Frz1-siRNA ϩ PIWI1a-CRISPR- activation ϩ MCAO, PIWI1a levels were sig- nificantlyincreasedwithanassociatedincrease in FOXM1 levels and a reduction in CC-3. These mechanistic studies support a link between PIWI1a and FOXM1. Studies have shown a significant asso- ciation between PIWIL1a and regulation of ␤-catenin (Reeves et al., 2012), but PIWIL1a has not been mechanistically linked with wnt3a or downstream signaling after stroke. We found that PIWI1a significantly decreased the levels of active p-GSK3␤ and significantly rescued levels of survival protein ␤-catenin. Inhibition of PIWI1a significantly increased levels of p-GSK3␤, suggesting that PIWI1a is an upstream regulator of GSK3␤. A link between Frz1 and p-GSK3␤ is supported by increased p-GSK3␤ secondary to Frz1- siRNA. To confirm PIWI1a’s role downstream of Frz1, PIWI1a was upregulated via CRISPR in combination with siRNA Frz1 Figure 11. Infarct volumes 24 h after pMCAO. A, Representative TTC staining images of coronal sections. B, wnt3a HD had no knockdown. This upregulation of PIWI1a sig- effect on infarct in the pMCAO compared with vehicle. C, Modified Garcia showed no improvement in neurological function after nificantly downregulated p-GSK3␤. wnt3a administration. D, Left forelimb placement showed no improvement after wnt3a treatment compared with vehicle. Each Frz1 has been reportedly activated by column represents the mean Ϯ SD (n ϭ 6/group). *p Ͻ 0.05 versus Sham. wnt3a and shown to be neuroprotective 6798 • J. Neurosci., July 25, 2018 • 38(30):6787–6801 Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO

Figure 13. Western blot confirmed the delivery of wnt3a in our treated animals. Right hemisphererepresentativeWesternblotsoftheratbrainandquantitativeanalysisshowedthat theexpressionofrecombinantGST-specifictaggedwnt3aproteinwassignificantlyhighercom- paredwithnontreatedgroups,bothshamandvehicle.EachcolumnrepresentsthemeanϮSD (n ϭ 6/group). *p Ͻ 0.05 versus Sham. #p Ͻ 0.05 versus vehicle.

Figure 12. Naive-treated animals showed a significant reduction of Frz-1 with Frz siRNA. A, Representative Western blots of the right hemisphere of the rat brain and quantitative analysis showed that the expression of Frz-1 significantly decreased at 24 h after Frz-siRNA administration compared with Naive and Naive ϩ Scrambled-siRNA (p Ͻ 0.05). B, PIWI1a levelsweresignificantlyattenuatedintheNaïveϩPIWI1asiRNAgroupcomparedtoNaïveand Naïve ϩ Scrambled siRNA groups. PIWI1a levels were restored to sham levels in the Naive ϩ PIWI1a-CRISPR ϩ Frz1-siRNA group. Each column represents the mean Ϯ SD (n ϭ 6/group). *p Ͻ 0.05 versus Naive. #p Ͻ 0.05 versus Naive ϩ Scrambled-siRNA. against A␤ oligomers (Chaco´n et al., 2008). We confirmed the presence of Frz1 in neurons using immunohistochemistry and Figure 14. LRP6 expression and activation after MCAO in rats. Right hemisphere represen- noticed an increased expression of Frz1 in the penumbra of the tativeWesternblotsoftheratbrainforLRP6,pLRP6,andactin.OverallLRP6expressionwasnot wnt3 ϩ MCAO group. Western blots supported that iN admin- significant between groups. However, phosphorylation of LRP significantly increased in the HD istration of wnt3a significantly increased the expression of Frz1 at treatment (1.2 ␮g/kg) treated groups, which was reversed in the Frz1-siRNA at 24HR after 24 and 72 h. These findings suggest a positive feedback loop MCAO. Each column represents the mean Ϯ SD (n ϭ 6/group). *p Ͻ 0.05 versus Sham. #p Ͻ involving wnt3a and Frz1. Additionally, the inhibition of PIWI1a 0.05 versus vehicle. &p Ͻ 0.05 versus wnt3a ϩ Scr siRNA. with siRNA resulted in decreased expression of Frz1, indicating that Frz1 may be transcriptionally regulated downstream of the Frz1 receptor (Ueno et al., 2013). The Frz1-positive feedback PIWI1a. Congruently, the canonical wnt3a/Frz1 pathway in can- loop requires further investigation. Inhibition of Frz1 downregu- cer tissues significantly decreased certain miRNAs (e.g., miR-204 lated wnt3a, PIWI1a, ␤-catenin, and FOXM1 while upregulating for Frz1), which may be a variable in the increased expression of p-GSK3␤ and CC-3 expression at 24 h. Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO J. Neurosci., July 25, 2018 • 38(30):6787–6801 • 6799

Interestingly, wnt3a protein expression was significantly re- ease of administration (Lin et al., 2009). We found that wnt3a’s duced in the Frz1-siRNA group. In support of this finding, delivery via an iN route was feasible and provided neuroprotec- miR-34 was observed to directly attenuate wnt3a protein expres- tion after stroke. Thus, iN administration of wnt3a is a clinically sion in breast cancer tissues (Song et al., 2015). In gastric carci- translatable route of administration, lowering risk of systemic noma cells, TGF-␤ induces caspase-8 activation and apoptosis by side effects. phosphorylating R-Smads, which form complexes with their One limitation of this study is that the design and results do common mediator, Smad4, and enter the nucleus to regulate not fully exclude the possibility of alternative pathways that act gene expression (Moustakas and Heldin, 2005). This Smad com- on wnt3a and downstream mediators; thus, further research will plex has been shown to attenuate the wnt expression in cancer cell need to investigate the relationship between PIWI1a, ␤-catenin, lines (Pang et al., 2011). A potential upstream link exists via the FOXM1, and the canonical apoptosis pathway to fully exclude or ability of GSK3␤ to phosphorylate and activate Smad4 to lower incorporate alternative pathways that may play a role in wnt3a the expression of wnt3a, an effect that is reversed after wnt3a physiology. For example, ␤-catenin canonically is known to treatment in HEK cells (Demagny et al., 2014). In light of several translocate into the nucleus and act as a transcriptional coactiva- promising hypotheses, further research is needed to elucidate the tor of transcription factors that belong to the TCF/LEF family gene regulation of wnt3a after Frz1 siRNA treatment in neurons (Rao and Ku¨hl, 2010). wnt3a is known to bind to Frz1a, expressed after MCAO in the rat model. Despite these limitations, we can by dendrites (Oderup et al., 2013) and microglia, which may play conclude that wnt3a exerts neuroprotective effects on PIWI1a/␤- a role in the regulation of apoptosis in neurons. Additionally, catenin/FOXM1 through the Frz1 receptor. GSK3␤ is part of the destruction complex (APC, Axin, and CK1) In the present study, we investigated two proapoptotic pro- (Minde et al., 2011)of␤-catenin and can be regulated by these teins, pGSK3␤ and CC-3. GSK-3␤ is reported to participate in intrinsic proteins, a process in which PIWI1a may be a significant neuron death through PI3K/AKT and wnt/␤-catenin signaling player. Further mechanistic studies of these three proteins inde- pathways after stroke (Chen et al., 2016). Also, after MCAO, pendently need to be investigated to fully understand the role of CC-3 significantly increased and caused apoptosis in neurons (Li PIWI1a in the wnt3a pathway. et al., 2017). Paralleling these findings, we observed an increase in In conclusion, iN administration of wnt3a was efficacious in pGSK3␤ and CC-3 at 24 and 72 h after MCAO. These effects were neuroprotection against transient cerebral ischemia in rats. New directly reversed with iN wnt3a treatment. This reduction in neu- pathway links between Frz1, PIWI1a, and FOXM1 have been ronal apoptotic proteins with wnt3a administration was associ- established by which wnt3a works in neurons to inhibit caspase- ated with significant neurobehavior improvement in forelimb 3-dependent apoptosis. A working model of these relationships is placement and recovery of modified Garcia scores to sham levels. shown in Figure 1. Given the lack of treatment options for isch- Studies have shown that MCAO causes deleterious neurological emic brain injury after stroke, these findings provide a basis for effects on modified Garcia (Wang et al., 2017) and left forelimb clinical trials to advance the clinical management of stroke and a placement (Senda et al., 2011), tests that correlate with cerebral foundation for future research in similar pathologies. wnt3a iN infarction. We examined the neuropathological damage at 24 and delivery should be investigated further as a potential therapeutic 72 h after MCAO and found our model to create significant neu- option for ischemic stroke. rological deficits. wnt3a caused a significant reduction of infarction volumes and neurological deficits in female rats. Efficacy of References wnt3a diminished with age, but significant benefits were seen in Abe T, Zhou P, Jackman K, Capone C, Casolla B, Hochrainer K, Kahles T, both infarct volume and forelimb placement. In a permanent Ross ME, Anrather J, Iadecola C (2013) Lipoprotein receptor-related occlusion model, there was no change after wnt3a administration protein-6 protects the brain from ischemic injury. Stroke 44:2284–2291. in infarct volume and neurological outcome. The proposed CrossRef Medline mechanism of action of wnt3a requires interaction of wnt3a with Arra´zola MS, Silva-Alvarez C, Inestrosa NC (2015) How the Wnt signaling pathway protects from neurodegeneration: the mitochondrial scenario. receptors, and thus direct sanguineous exposure to the area of Front Cell Neurosci 9:166. CrossRef Medline infarction is likely required for therapeutic effect. While the pen- Ayer RE, Jafarian N, Chen W, Applegate RL 2nd, Colohan AR, Zhang JH umbra of the infarction zone receives some perfusion and thus (2012) Preoperative mucosal tolerance to brain antigens and a neuropro- wnt3a exposure in models of permanent occlusion, without rep- tective immune response following surgical brain injury. J Neurosurg erfusion, there is limited opportunity for rescue of severely hy- 116:246–253. CrossRef Medline poxic penumbra. In agreement with our findings, a significant Bhanot P, Brink M, Samos CH, Hsieh JC, Wang Y, Macke JP, Andrew D, Nathans J, Nusse R (1996) A new member of the frizzled family from number of proposed neuroprotective agents have only demon- Drosophila functions as a wingless receptor. Nature 382:225–230. CrossRef strated effectiveness in animal models of transient occlusion Medline (Min et al., 2013). Although wnt3a administered restored neuro- Bortesi L, Fischer R (2015) The CRISPR/Cas9 system for plant genome ed- logical testing scores to sham levels in treatment groups, significance iting and beyond. Biotechnol Adv 33:41–52. CrossRef Medline was only seen versus vehicle groups in left forelimb placement tests. Bourhis E, Tam C, Franke Y, Bazan JF, Ernst J, Hwang J, Costa M, Cochran While infarction volume is known to affect neurobehavior tests, AG, Hannoush RN (2010) Reconstitution of a frizzled8.wnt3a.LRP6 signaling complex reveals multiple wnt and Dkk1 binding sites on LRP6. these tests are not individually optimal to evaluate the total dam- J Biol Chem 285:9172–9179. CrossRef Medline age in the stroke because each test is sensitive to specific focal Candelario-Jalil E (2009) Injury and repair mechanisms in ischemic stroke: neurological deficits (Senda et al., 2011); nonetheless, this treat- considerations for the development of novel neurotherapeutics. Curr ment shows promise in stroke rehabilitation. Opin Investig Drugs 10:644–654. Medline Treatments targeting the CNS that are administered outside Cerpa W, Toledo EM, Varela-Nallar L, Inestrosa NC (2009) The role of wnt of the CNS must be able to efficiently cross the blood–brain signaling in neuroprotection. Drug News Perspect 22:579–591. CrossRef Medline barrier (Thorne and Frey, 2001). iN administration of neuropro- Chaco´n MA, Varela-Nallar L, Inestrosa NC (2008) Frizzled-1 is involved in tectants has previously been established as a viable route of ad- the neuroprotective effect of wnt3a against abeta oligomers. J Cell Physiol ministration for stroke, allowing rapid delivery to the CNS and 217:215–227. CrossRef Medline 6800 • J. Neurosci., July 25, 2018 • 38(30):6787–6801 Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO

Chen C, Ostrowski RP, Zhou C, Tang J, Zhang JH (2010) Suppression of Lin S, Fan LW, Rhodes PG, Cai Z (2009) Intranasal administration of IGF-1 hypoxia-inducible factor-1alpha and its downstream genes reduces acute attenuates hypoxic-ischemic brain injury in neonatal rats. Exp Neurol hyperglycemia-enhanced hemorrhagic transformation in a rat model of 217:361–370. CrossRef Medline cerebral ischemia. J Neurosci Res 88:2046–2055. CrossRef Medline Liu L, Puri KD, Penninger JM, Kubes P (2007) Leukocyte PI3Kgamma and Chen S, Ma Q, Krafft PR, Hu Q, Rolland W 2nd, Sherchan P, Zhang J, Tang J, PI3K␦ have temporally distinct roles for leukocyte recruitment in vivo. Zhang JH (2013) P2X7R/cryopyrin inflammasome axis inhibition re- Blood 110:1191–1198. CrossRef Medline duces neuroinflammation after SAH. Neurobiol Dis 58:296–307. CrossRef Li C, Che LH, Ji TF, Shi L, Yu JL (2017) Effects of the TLR4 signaling Medline pathway on apoptosis of neuronal cells in diabetes mellitus compli- Chen W, Ma Q, Suzuki H, Hartman R, Tang J, Zhang JH (2011) Osteopon- cated with cerebral infarction in a rat model. Sci Rep 7:srep43834. tin reduced hypoxia-ischemia neonatal brain injury by suppression of CrossRef Medline apoptosis in a rat pup model. Stroke 42:764–769. CrossRef Medline MacDonald BT, Tamai K, He X (2009) Wnt/␤-catenin signaling: compo- Chen Y, Zhang Y, Tang J, Liu F, Hu Q, Luo C, Tang J, Feng H, Zhang JH nents, mechanisms, and diseases. Dev Cell 17:9–26. CrossRef Medline ␤ (2015) Norrin protected blood–brain barrier via frizzled-4/ -catenin Mackay J, Mensah G (2004) Atlas of heart disease and stroke, Ed 1. Geneva: pathway after subarachnoid hemorrhage in rats. Stroke 46:529–536. World Health Organization. CrossRef Medline Mali P, Yang L, Esvelt KM, Aach J, Guell M, DiCarlo JE, Norville JE, Church Chen X, Liu Y, Zhu J, Lei S, Dong Y, Li L, Jiang B, Tan L, Wu J, Yu S, Zhao Y GM (2013) RNA-guided human genome engineering via Cas9. Science (2016) GSK-3␤ downregulates Nrf2 in cultured cortical neurons and in a 339:823–826. CrossRef Medline rat model of cerebral ischemia-reperfusion. Sci Rep 6:20196. CrossRef McBride DW, Tang J, Zhang JH (2016) Development of an infarct volume Medline algorithm to correct for brain swelling after ischemic stroke in rats. Acta Cohen P, Goedert M (2004) GSK3 inhibitors: development and therapeutic Neurochir Suppl 121:103–109. CrossRef Medline potential. Nat Rev Drug Discov 3:479–487. CrossRef Medline Michalski D, Heindl M, Kacza J, Laignel F, Kappers-Tiedt L, Schneider D, Collino M, Aragno M, Castiglia S, Tomasinelli C, Thiemermann C, Boccuzzi G, Fantozzi R (2009) Insulin reduces cerebral ischemia/reperfusion in- Grosche J, Boltze J, Lahr M, Hobohm C, Hartig W (2012) Spatio- jury in the hippocampus of diabetic rats. Diabetes 58:235–242. CrossRef temporal course of macrophage-like cell accumulation after experimental Medline embolic stroke depending on treatment with tissue plasminogen activator Demagny H, Araki T, De Robertis EM (2014) The tumor suppressor and its combination with hyperbaric oxygenation. Eur J Histochem 56:14. Smad4/DPC4 is regulated by phosphorylations that integrate FGF, wnt, CrossRef Medline and TGF-␤ signaling. Cell Rep 9:688–700. CrossRef Medline Min J, Farooq MU, Gorelick PB (2013) Neuroprotective agents in ischemic Ferrer I, Planas AM (2003) Signaling of cell death and cell survival following stroke: past failures and future opportunities. Clin Invest 3:1167–1177. focal cerebral ischemia: life and death struggle in the penumbra. J Neuro- CrossRef pathol Exp Neurol 62:329–339. CrossRef Medline Minde DP, Anvarian Z, Ru¨diger SG, Maurice MM (2011) Messing up dis- Fisher M, Brott TG (2003) Emerging therapies for acute ischemic stroke order: how do missense mutations in the tumor suppressor protein APC new therapies on trial. Stroke 34:359–361. CrossRef Medline lead to cancer? Mol Cancer 10:101. CrossRef Medline Gratz SJ, Cummings AM, Nguyen JN, Hamm DC, Donohue LK, Harrison Moustakas A, Heldin CH (2005) Non-Smad TGF-␤ signals. J Cell Sci 118: MM, Wildonger J, O’Connor-Giles KM (2013) Genome engineering of 3573–3584. CrossRef Medline Drosophila with the CRISPR RNA-guided Cas9 nuclease. Genetics 194: Oderup C, LaJevic M, Butcher EC (2013) Canonical and noncanonical Wnt 1029–1035. CrossRef Medline proteins program dendritic cell responses for tolerance. J Immunol 190: Horvath P, Barrangou R (2010) CRISPR/Cas, the immune system of bacte- 6126–6134. CrossRef Medline ria and archaea. Science 327:167–170. CrossRef Medline Pang L, Qiu T, Cao X, Wan M (2011) Apoptotic role of TGF-␤ mediated by Hu Q, Chen C, Yan J, Yang X, Shi X, Zhao J, Lei J, Yang L, Wang K, Chen L, Smad4 mitochondria translocation and cytochrome c oxidase subunit II Huang H, Han J, Zhang JH, Zhou C (2009) Therapeutic application of interaction. Exp Cell Res 317:1608–1620. CrossRef Medline gene silencing MMP-9 in a middle cerebral artery occlusion-induced focal Perez-Pinera P, Kocak DD, Vockley CM, Adler AF, Kabadi AM, Polstein LR, ischemia rat model. Exp Neurol 216:35–46. CrossRef Medline Thakore PI, Glass KA, Ousterout DG, Leong KW, Guilak F, Crawford GE, Inestrosa NC, Varela-Nallar L (2014) Wnt signaling in the nervous system Reddy TE, Gersbach CA (2013) RNA-guided gene activation by CRISPR- and in Alzheimer’s disease. J Mol Cell Biol 6:64–74. CrossRef Medline Cas9-based transcription factors. Nat Methods 10:973–976. CrossRef Jiang L, Wang P, Chen L, Chen H (2014) Downregulation of FoxM1 by Medline thiostrepton or small interfering RNA inhibits proliferation, transforma- R&D Systems (2018) Wnt receptors and pathways. https://www.rndsystems. tion ability and angiogenesis, and induces apoptosis of nasopharyngeal com/resources/articles/wnt-receptors-pathways. carcinoma cells. Int J Clin Exp Pathol 7:5450–5460. Medline Rao TP, Ku¨hl M (2010) An updated overview on wnt signaling pathways: a Jiang W, Bikard D, Cox D, Zhang F, Marraffini LA (2013) RNA-guided prelude for more. Circ Res 106:1798–1806. CrossRef Medline editing of bacterial genomes using CRISPR-cas systems. Nat Biotechnol Reeves ME, Baldwin ML, Aragon R, Baldwin S, Chen ST, Li X, Mohan S, 31:233–239. CrossRef Medline Amaar YG (2012) RASSF1C modulates the expression of a stem cell Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E (2012) renewal gene, PIWIL1. BMC Res Notes 5:239. CrossRef Medline A programmable dual-RNA-guided DNA endonuclease in adaptive bac- Rha JH, Saver JL (2007) The impact of recanalization on ischemic stroke terial immunity. Science 337:816–821. CrossRef Medline outcome: a meta-analysis. Stroke 38:967–973. CrossRef Medline Kawabori M, Kacimi R, Kauppinen T, Calosing C, Kim JY, Hsieh CL, Naka- Roth JM (2011) Recombinant tissue plasminogen activator for the treat- mura MC, Yenari MA (2015) Triggering receptor expressed on myeloid cells 2 (TREM2) deficiency attenuates phagocytic activities of microglia ment of acute ischemic stroke. Proc Bayl Univ Med Cent 24:257–259. and exacerbates ischemic damage in experimental stroke. J Neurosci 35: CrossRef Medline 3384–3396. CrossRef Medline Schmued LC, Stowers CC, Scallet AC, Xu L (2005) Fluoro-jade C results in Kelly S, Zhao H, Hua Sun G, Cheng D, Qiao Y, Luo J, Martin K, Steinberg GK, ultra high resolution and contrast labeling of degenerating neurons. Brain Harrison SD, Yenari MA (2004) Glycogen synthase kinase 3␤ inhibitor Res 1035:24–31. CrossRef Medline Chir025 reduces neuronal death resulting from oxygen-glucose depriva- Senda DM, Franzin S, Mori MA, de Oliveira RM, Milani H (2011) Acute, tion, glutamate excitotoxicity, and cerebral ischemia. Exp Neurol 188: post-ischemic sensorimotor deficits correlate positively with infarct size 378–386. CrossRef Medline but fail to predict its occurrence and magnitude after middle cerebral Kusaka I, Kusaka G, Zhou C, Ishikawa M, Nanda A, Granger DN, Zhang JH, artery occlusion in rats. Behav Brain Res 216:29–35. CrossRef Medline Tang J (2004) Role of AT1 receptors and NAD(P)H oxidase in diabetes- Shruster A, Ben-Zur T, Melamed E, Offen D (2012) Wnt signaling enhances aggravated ischemic brain injury. Am J Physiol Heart Circ Physiol 286: neurogenesis and improves neurological function after focal ischemic H2442–H2451. CrossRef Medline injury. PLoS One 7:e40843. CrossRef Medline Lai TW, Zhang S, Wang YT (2014) Excitotoxicity and stroke: identifying Song JL, Nigam P, Tektas SS, Selva E (2015) microRNA regulation of wnt novel targets for neuroprotection. Prog Neurobiol 115:157–188. CrossRef signaling pathways in development and disease. Cell Signal 27:1380– Medline 1391. CrossRef Medline Matei et al. • wnt3a Attenuates Neuronal Apoptosis after MCAO J. Neurosci., July 25, 2018 • 38(30):6787–6801 • 6801

Taylor RA, Sansing LH (2013) Microglial responses after ischemic Wu X, Deng G, Hao X, Li Y, Zeng J, Ma C, He Y, Liu X, Wang Y (2014) A stroke and intracerebral hemorrhage. J Immunol Res 2013:e746068. caspase-dependent pathway is involved in Wnt/␤-catenin signaling pro- CrossRef Medline moted apoptosis in Bacillus Calmette-Guerin infected RAW264.7 macro- Thorne RG, Frey WH 2nd (2001) Delivery of neurotrophic factors to the phages. Int J Mol Sci 15:5045–5062. CrossRef Medline central nervous system: pharmacokinetic considerations. Clin Pharmaco- Zhang N, Wei P, Gong A, Chiu WT, Lee HT, Colman H, Huang H, Xue J, Liu kinet 40:907–946. CrossRef Medline M, Wang Y, Sawaya R, Xie K, Yung WK, Medema RH, He X, Huang S Ueno K, Hirata H, Hinoda Y, Dahiya R (2013) Frizzled homolog proteins, (2011) FoxM1 promotes ␤-catenin nuclear localization and controls wnt microRNAs and wnt signaling in cancer. Int J Cancer 132:1731–1740. target-gene expression and glioma tumorigenesis. Cancer Cell 20:427– CrossRef Medline 442. CrossRef Medline Wang Y, Sherchan P, Huang L, Akyol O, McBride DW, Zhang JH (2017) Zhao Y, Wei ZZ, Zhang JY, Zhang Y, Won S, Sun J, Yu SP, Li J, Wei L (2017) Naja sputatrix venom preconditioning attenuates neuroinflammation in GSK-3␤ inhibition induced neuroprotection, regeneration, and func- a rat model of surgical brain injury via PLA2/5-LOX/LTB4 cascade acti- tional recovery after intracerebral hemorrhagic stroke. Cell Transplant vation. Sci Rep 7:5466. CrossRef Medline 26:395–407. CrossRef Medline